Scenario Planning: Facilitating the 2x2 Critical Uncertainties Workshop

Most strategy teams do not fail because they chose the wrong future — they fail because they only imagined one. The 2x2 critical uncertainties workshop is the discipline that forces you to hold four futures simultaneously and build strategy robust enough to survive all of them.

This is a facilitation guide for running that workshop. It covers every phase: defining the focal question, scanning and ranking drivers, selecting axes, building narratives, and stress-testing strategy. It assumes you are facilitating a real team with real decisions to make, not running a futures literacy exercise for its own sake.

What the 2x2 method is and why it holds up

The 2x2 scenario matrix was developed by Shell's Group Planning team in the 1970s, led by Pierre Wack. When the OPEC embargo hit in 1973, Shell was the only major oil company with contingency strategies already in place. Within two years it moved from seventh to second in industry profitability. The method has been documented extensively by Peter Schwartz in The Art of the Long View and refined by GBN, Reos Partners, and the Oxford Futures Library over five decades.

What makes it durable is what it does to cognition. Different participants hold different mental models of the future. Left unstructured, groups converge on whoever speaks most confidently. The 2x2 forces productive conflict: you cannot build four genuinely distinct futures without surfacing the disagreements that would otherwise stay polite and implicit. This is what Daniel Kahneman calls "outside view" thinking — deliberately stepping outside your own assumptions to reason about the range of possible states rather than the most comfortable one.

The method is specifically designed for high-ambiguity environments where historical data is a weak guide. If you can forecast it reliably, you do not need scenario planning. If you cannot, you do.

Phase 1: Defining the focal question

This is where most workshops go wrong before they begin. Teams arrive wanting to explore "the future of our industry" or "what the world will look like in 2035." Those are not focal questions. They produce scenarios that feel intellectually interesting and are disconnected from any real decision.

A focal question must be framed as a decision: Should we invest in manufacturing capacity in Southeast Asia by 2030? or How should we allocate R&D budget across our three product lines given uncertainty about regulatory and market conditions through 2032? The test is simple: if you cannot explain how each scenario changes what you would actually do, the focal question is too broad.

Time horizon matters too. As Peter Schwartz has outlined through GBN, the horizon should be long enough that today's constraints feel breakable but short enough that participants can still reason about plausible cause-and-effect chains. For most industries, 7-10 years works. For fast-moving technology sectors, 3-5 years is more appropriate.

Before the workshop, spend 30-45 minutes in stakeholder interviews surfacing the "taken-for-granted" assumptions your leadership team holds about the focal question. These assumptions are often the most strategically dangerous blind spots, and making them explicit before the room fills with people gives you a significant facilitation advantage.

When PATH, the global health organization, used scenario planning to examine vaccine delivery in low-income countries, its focal question was tightly scoped: How should PATH prioritize its cold-chain investment strategy given uncertainty about infrastructure investment and local manufacturing capacity by 2030? That specificity meant each scenario directly tested real capital allocation decisions rather than producing abstract narratives. That is the standard to aim for.

Phase 2: The STEEP-V driver scan and ranking

Before you can define axes, you need the full landscape of forces shaping the focal question on the table. The STEEP-V framework — Social, Technological, Economic, Environmental, Political, Values — gives participants a structured scaffold for generating drivers without premature filtering. Run 20-30 minutes of individual sticky-note generation followed by affinity clustering. Expect 40-80 candidate drivers. Do not filter during generation.

Once the drivers are visible, each one gets evaluated on two independent dimensions: its impact on the focal question, and the degree of uncertainty around how it will play out. Drivers that are high-impact but low-uncertainty become "predetermined elements" — the backdrop facts assumed true across all four scenarios (global population aging, for instance, or the irreversibility of certain climate commitments). Only drivers that are both high-impact and high-uncertainty are candidates for your axes.

This distinction, drawn from Kees van der Heijden's published facilitation frameworks in Scenarios: The Art of Strategic Conversation, is the most commonly skipped step and the most consequential. Skip it and you end up with axes built on things that are actually quite predictable, which makes your scenarios false uncertainty exercises.

Display the impact/uncertainty matrix on a large wall. Place drivers on sticky notes. Let the full group see and contest placements in real time. The visible artefact is not decorative — it surfaces disagreements that would otherwise remain implicit. Your goal coming out of this phase is a shortlist of 6-8 high-impact, high-uncertainty drivers to bring into axis selection.

The OECD's Strategic Foresight work on the future of higher education provides a useful reference here. Facilitators ran a STEEP scan across 14 member country representatives, clustered roughly 60 drivers, and identified "pace of automation displacing graduate-level jobs" and "trust in institutional credentials" as the two axis candidates — both genuinely high-impact and genuinely uncertain. Those axes produced scenarios that directly informed UNESCO and national ministry policy debates.

Phase 3: Selecting the axes

This is the highest-leverage conversation in the workshop. It is also the most likely to go badly.

Axes must meet three criteria simultaneously. Each axis must span a range of plausible outcomes, not just a spectrum from optimistic to pessimistic. The two axes must be genuinely independent — all four quadrants must be logically possible, not just two or three. And the combination must be maximally relevant to the focal question.

The independence test is the one facilitators most often fail to apply rigorously. Ask the group: "Can we tell a plausible story where the high end of Axis A occurs alongside the low end of Axis B?" Do this for all four quadrant combinations. If the answer is no for any combination, the axes are correlated and you will end up with effectively two distinct futures dressed as four. The World Economic Forum's Global Risks scenario work has repeatedly used "degree of multilateral cooperation" and "pace of technological disruption" as axes precisely because they pass this test — the high-cooperation/slow-tech world and the low-cooperation/fast-tech world are both genuinely plausible.

The naming of poles matters more than most facilitators expect. "Distributed Power / Concentrated Power" is generative. "Fragmented Chaos / Strong Leadership" is not — it embeds value judgments that constrain participants' imagination in the narrative phase. Keep pole names neutral and evocative.

Phase 4: Building the four scenario narratives

Split into breakout groups of 3-5 people, one per quadrant. Give each group a structured template:

• What sequence of events led to this world?
• Who are the winners and losers?
• What does our organization look like in this world?
• What headline from a major publication captures this future?

The GBN approach of writing a two-paragraph "newspaper front page" story set at the end of the time horizon is worth using. It activates narrative cognition rather than analytical cognition, which helps participants genuinely inhabit the future rather than describe it from a safe distance.

When groups present back, your job as facilitator is to check that the four narratives are genuinely different. It is common for groups to unconsciously hedge toward the present, producing scenarios that feel like mild variations on today. Push them: if you cannot tell these four stories apart at a dinner party, they are not distinct enough.

Explicitly normalize the worst-case quadrant. Research by Harvard's Amy Edmondson on psychological safety shows that groups avoid exploring unfavorable scenarios when the social cost of being the person who imagines failure feels high. Teams that soften worst-case quadrants systematically understate downside risk. The facilitator needs to create permission to go there.

The Mont Fleur scenarios from South Africa (1991-1992) remain the canonical example of what good narrative work produces. Facilitated by Adam Kahane with a cross-partisan group of 22 South Africans, the four scenarios — including "Flamingo" (slow sustainable flight) and "Icarus" (unsustainable populist spending) — had vivid names and story arcs that were distributed via video across the country. South African policymakers credited the scenarios with broadening the range of acceptable policy debate during constitutional negotiations. Narrative did what analytical frameworks could not.

Phase 5: Stress-testing strategy against the scenarios

Once narratives are complete, the workshop shifts from futures imagination to strategic assessment. For each existing or proposed strategic option, run a robustness check across all four quadrants: does this option perform well, adequately, or poorly in each scenario?

Options that perform well across all four quadrants are your robust strategies — the highest-priority investments regardless of which future materializes. A second category covers options that perform well in high-probability scenarios but limit downside in the worst case. These are hedging strategies, deployed conditionally. A third category covers "big bets" — options that only make sense if leading indicators suggest the world is moving toward a specific quadrant.

This three-tier framing prevents the common mistake of treating all strategic options as binary go/no-go decisions. Unilever's Sustainable Living Plan, developed partly through scenario stress-testing with Forum for the Future, illustrates the approach. Facilitators tested Unilever's core business model against scenarios ranging from strong regulatory pressure for sustainability to consumer backlash against green premiums. Strategies that performed well across all scenarios — including reformulating products to reduce water use regardless of regulatory requirements — became foundational commitments.

Close the stress-test step by identifying early warning indicators: specific, measurable signals that would suggest the world is moving toward one quadrant rather than others. Assign ownership. Set a review cadence. Without this step, the scenario work stays a workshop output. With it, the workshop becomes the beginning of an ongoing strategic intelligence practice.

Facilitation design: timing, roles, and what goes wrong

A full 2x2 workshop requires 6-8 hours of facilitated time for 8-20 participants. You can split it across two half-day sessions with overnight reflection between the driver ranking and narrative construction phases — that gap is genuinely useful for synthesis. Compressing below four hours consistently produces axis choices that participants have not internalized. They will not own the scenarios, and the scenarios will not survive contact with the next budget conversation.

The UK Government Office for Science Futures Toolkit specifies a role separation worth adopting: a lead facilitator focused on process, a domain expert available for content questions, and a dedicated recorder. This prevents the most common facilitation failure mode — the facilitator getting pulled into content debates and losing process control.

The three failure modes that appear most consistently across GBN, Reos Partners, and Oxford Futures Library practitioner documentation:

• Axes that are correlated rather than independent, producing only two effective scenarios dressed as four
• Narrative groups that default to trend extrapolation, making all four quadrants feel like slight variations on the present
• No early warning indicators identified, leaving the workshop with no mechanism to keep scenarios alive in ongoing management conversations

Workshop Weaver has facilitation templates for each phase of this process, including the driver ranking matrix, axis selection criteria, narrative templates, and the strategy robustness grid. If you are running this for the first time, working from a structured template is significantly better than improvising the artefacts on the day.

Running your first session

The 2x2 scenario workshop is not a one-time event. Done well, it produces four futures the team owns, a set of robust strategic choices, a list of big bets with trigger conditions, and a scanning programme to track which future is emerging. That is the beginning of a continuous strategic conversation, not the end of one.

To go deeper on the method itself, the scenario planning 2x2 method page covers the theoretical foundations in detail. For reference on the critical uncertainties framework specifically, see the critical uncertainties method page. And if you are thinking about how this workshop fits into a broader strategy offsite, the strategy workshop design guide covers sequencing and stakeholder preparation.

If you want facilitation support or a ready-to-run template, book a workshop design consultation through Workshop Weaver or download the scenario planning template directly from the site. The template includes the STEEP-V scan, the impact/uncertainty ranking matrix, the axis selection criteria checklist, the narrative briefing sheet, and the strategy robustness grid — everything you need to run the session without building artefacts from scratch.